Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I have a computer running 2 256GB Crucial M4's (CT256M4SSD2) in a RAID 0 (striped) array on an ASUS P9X79 Pro using Intel's (built-in) RAID system.

I recently installed Windows 8 Pro as UEFI on this RAID array. (wiping a fully-function Windows 7 non-UEFI installation)

Now, whenever the computer is left running for about an hour, the system no longer sees those drives. Since those drives contain Windows, this leads to various forms of BSODs. If I Intel RSTe (RAID manager) is running at the time, it will say that the disk backing that RAID array has been removed.

Once this happens, if I reset the computer, it will no longer boot. Entering BIOS setup shows that the SATA 3 (6Gbps) ports that those disks are connected to are both empty.

If I then power down the system completely, then turn it on again, the drives reappear, but the problem repeats after another hour or so.

I have inconclusively determined that the problem occurs even if Windows is not running (booted into the installation environment from a UEFI flash drive)

I don't think there has been any data corruption since this started happening, although I have had two strange issues with a GIT repo on that disk.
sfc /ScanNow and Intel's disk check (in RSTe) both do not find anything.

Does anyone know what might cause this?

share|improve this question
3  
Whatever it is, you can be pretty sure it wasn't Windows 8. This is clearly a hardware issue. –  Michael Hampton Oct 31 '12 at 22:08
1  
I've had countless errors with Intel's RST raid "losing" SSD drives over the past year with Windows 7. It's a hardware issue in their controller it seems like, as they don't even show up in the BIOS despite being plugged in, and work just fine in other computers. –  Darth Android Oct 31 '12 at 22:09
    
This sounds exactly like the firmware bug that was patched in January. See my answer for more details. –  rob Oct 31 '12 at 22:23
    
@MichaelHampton: Yes; I'm pretty sure it's a hardware issue, but it's still a very strange coincidence. –  SLaks Oct 31 '12 at 22:41
    
I've seen this happen with encryption -- e.g., SecureDoc. Are you running full-disk or other encryption? –  belacqua Oct 31 '12 at 22:46

2 Answers 2

up vote 4 down vote accepted

Make sure you're running the latest firmware on your Crucial M4s. There is a well-known firmware bug that causes the drives to misbehave (causing BSODs in Windows) every hour or so, once the drives have hit 5184 power-on hours.

You can use a SMART diagnostic such as CrystalDiskInfo or smartmontools to check the power-on hours and see if you recently passed the magic number.

Crucial patched this bug in a January 2012 firmware update.  The update leaves your data intact, but you should back up anything important first, even if it requires several 1-hour sessions and you have to do it in chunks.

share|improve this answer
    
I'll try this now. Fortunately, I already have full backups, from both before and after Windows 8. I think these disks what have had a little bit more than 5184 hours of uptime. –  SLaks Oct 31 '12 at 22:34
1  
You're right; CDI says 5256 hours, and this started happening at the beginning of this week. I will hopefully accept this answer in an hour. :) –  SLaks Oct 31 '12 at 22:56
    
@SLaks: Directly related: http://superuser.com/q/467659 –  Deltik Oct 31 '12 at 23:27
    
Everything works fine now. Thank you very much. –  SLaks Nov 1 '12 at 2:22

I contacted Crucial tech support and got the following reply: (8 hours later)

Thank you for contacting us. The behavior you are describing is consistent with TRIM commands not executing on the drive (which may result from lots of writing and rewriting, especially OS installs, not done on the OS level). The accumulation of 'junk' data from file deletions, even reformats, basically clogs up the drive and reduces or sometimes even halts performance and can even dismount the drive. A secondary feature called Garbage Collection activates when the drive is powered, but has no data throughput, for an extended period, and does background cleanup on our SSDs which can make up for the lack of TRIM.

If this is the cause of your drive's behavior, a period of idling the drive without any data being written or read actively will improve its performance. We recommend you perform this idle period on a desktop computer because it allows you to only connect the SATA power connection. However, a USB enclosure with an external power source will also work. A laptop computer will also work, but you’ll have to connect the drive and navigate to the systems BIOS menu. (Please refer to your system manufacturer’s documentation on how to access the BIOS.)

For laptops, we don’t recommended using a USB enclosure powered via USB; dedicated AC power is preferable. In addition, Apple users must hold the Option key while they power on the system (with the SSD installed). This will boot the Mac to the Startup Manager screen. The Startup Manager screen works like the BIOS screen on a laptop, in that it gives the drive power without any data throughput. An alternative Apple method is to close all open programs (including background processes) after startup and idle on your desktop. Just power the drive and let it sit, preferably overnight, but even a few hours should improve performance if this is the cause of your drive's erratic performance, though full Garbage Collection takes 8-24 hours and should be resumed when you can be without that system for an extended period.

This was not the problem (rob's answer fixed the problem immediately), but it may help someone else.

I was using Intel's x79 RAID controller, which does send TRIM commands, so I don't think this was applicable to me.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.